Managed ClickHouse on AWS

Free credit available Limited time

Managed ClickHouse® by Selfhost.dev, real-time analytics, run for you.

Selfhost.dev runs a dedicated, columnar OLAP database on AWS for sub-second analytics over billions of rows: events, logs and dashboards. Keeper-based high availability, HTTP and native protocols, SQL users and automated backups, all managed for you. Bring your own AWS account or run it in ours and manage it from your AI editor.

Deploy ClickHouse See pricing
ClickHouse 24.3, 24.8, 25.1 / AWS, worldwide / From ~$27/mo / Pay-as-you-go
Powered by AWS

The reliable home for your data.

Amazon Web Services (AWS)
Dedicated instances, run for you

Dedicated PostgreSQL, MySQL, Redis and ClickHouse on AWS, with BYOC into your own account. Multi-AZ on all four, PITR on Postgres and MySQL.

In our AWS account or your own with BYOC. You get AWS uptime and performance; we handle the ops.

What you get

Analytics-grade ClickHouse.
Without running the cluster yourself.

Instant provisioning

A 5-step wizard: pick ClickHouse 24.3, 24.8 or 25.1, cloud and region, instance size and storage, backup policy, review and deploy. Live in minutes.

Columnar OLAP engine

Purpose-built for analytics: sub-second aggregations across billions of rows, high-throughput inserts and queries that would crawl on a row store. Real-time, not batch.

EBS, S3 or tiered storage

By default your MergeTree data lives in S3-compatible object storage (S3, MinIO or R2) on a bucket we provision for you. Pick EBS-only instead or tiered: hot data on EBS with cold data aged into object storage by TTL. Backup and restore work in every mode.

HTTP and native protocols

The HTTP API on port 8123 for JDBC and most clients, plus the native TCP protocol on 9000 for high-throughput ingest. Connect from clickhouse-client or anything that speaks either.

Keeper high availability

Optional: turn on Multi-AZ for a 3-node cluster with embedded ClickHouse Keeper (Raft quorum) and automatic failover, plus distributed queries across the selfhost_cluster. Production resilience without wiring up ZooKeeper.

Replication and read scale

Add replicas for read throughput and redundancy. All nodes accept reads, so heavy analytics fan out across the cluster instead of hammering one box.

Backups, snapshots, fork

Manual snapshots plus automated backup policies with the retention you choose and fork an independent copy of a dataset for testing or a one-off analysis without touching production.

Autoscaling, scheduled or reactive

Resize on demand as your data grows, set reactive autoscaling rules that follow load or schedule start and stop windows so an analytics box that only runs during business hours costs you only then. Storage ships with 3,000 IOPS included free.

SQL users and monitoring

Database users with roles, plus live metrics (queries per second, rows inserted per second, merges, replication delay) and email and Slack alert rules on the thresholds you care about.

Locked down by default

IP whitelisting, delete protection and public access only if you turn it on. Your analytics database is not on the open internet by accident.

Encrypted in transit and at rest

Connections use TLS. Data at rest is encrypted on EBS with AWS KMS keys and in S3-backed object storage with S3 server-side encryption (AES256), so snapshots and backups are encrypted too.

Security and trust

Your data: encrypted, isolated,
and always under your control.

Security is not a separate tier here. Every managed instance ships locked down by default, encrypted end to end and yours to export or run in your own cloud.

See how we secure and run your data

Encrypted in transit and at rest

TLS on the wire, AWS KMS on disk. Snapshots and backups are encrypted too.

Single-tenant, dedicated instance

Your own dedicated instance, not a shared multi-tenant database pool.

Private by default

IP whitelisting and delete protection are on. Public access only if you turn it on.

Multi-AZ, PITR and backups

Automatic failover on every engine, point-in-time restore on PostgreSQL and MySQL, automated backups on all.

Your data stays yours (BYOC)

Run it in your own AWS account. Standard engines, nothing proprietary, yours to keep.

Live system status Platform availability from our public status page. View live status

Built for analytics

When Postgres starts to sweat
on your analytics.

Product and event analytics

Clickstream, funnels, retention and usage over billions of events, queried in real time instead of an overnight job.

Logs and observability

High-volume logs, metrics and traces stored cheaply and searched fast, a self-hosted backbone for your observability stack.

Real-time dashboards

Sub-second aggregations that power live BI dashboards and in-app charts without a caching layer bolted in front.

Time-series at scale

IoT readings, financial ticks and any firehose of timestamped data that a row store chokes on past a certain volume. See how it compares to purpose-built engines in our best time-series database guide.

ClickHouse is an OLAP engine, built for reads and aggregations, not transactions. For the app's own writes and relational data, pair it with managed PostgreSQL or MySQL on the same account and stream events into ClickHouse for the analytics side.

A managed ClickHouse Cloud alternative

A dedicated instance
you can actually see into.

A real dedicated ClickHouse instance on AWS, not a black-box serverless tier. You choose the instance size, see live metrics and logs, keep it in your own AWS account with BYOC and pay by the hour with no seat or tier minimums. Everything pauses at a zero balance, so an analytics box does not quietly bill you between reports. On a like-for-like always-on instance (2 vCPU, 8 GB RAM, 100 GB) that works out to about $65/mo here versus about $173/mo on ClickHouse Cloud. See the full comparison, our roundup of the best managed ClickHouse services or how it compares with other real-time analytics databases.

  • ✓ Dedicated instance, your sizing, no noisy neighbors
  • ✓ Pay-as-you-go credits, no tier or subscription minimum
  • ✓ Run it in your own AWS account with BYOC

Moving off ClickHouse Cloud? A Selfhost.dev engineer plans and runs the migration with you, free on every plan size. See how it works.

Deploy managed ClickHouse
Transparent pricing
BillingBy the hour, prepaid credits
Tiers and subscriptionsNone
Entry instancefrom ~$27/mo

ClickHouse wants memory headroom, so instances start at 4 GB RAM. A public IPv4 (about $5/mo) is added at deploy. Everything pauses at a zero balance, so there is never a surprise invoice. Build your exact bill.

Bring your own cloud

Our management.
Your AWS account.

Run managed ClickHouse inside your own AWS account with BYOC. You keep full ownership of the infrastructure, the VPC and the data residency, while Selfhost.dev handles provisioning, backups, failover and upgrades. The fee is 70% of the managed price and you pay AWS directly for the compute and the storage. New to the model? Here's what BYOC is and when it makes sense.

  • ✓ Connect via access keys or an IAM role
  • ✓ Your VPC, your security groups, your region
  • ✓ No enterprise contract required
Same data, your account
ComputeYour EC2
Storage and snapshotsYour EBS
Data residencyYour region
Our fee70% of managed price

If you ever leave, the database stays in your account. No lock-in.

AI-native management

Manage ClickHouse with AI.
Right from your editor.

Provision instances, run backups, add replicas, configure alerts and inspect performance from Claude, Cursor, Windsurf or any MCP client. 189 MCP tools across the platform, every one mapped to a real API call, with destructive operations gated behind explicit confirmation.

Explore the MCP server

Frequently asked questions

What is Selfhost.dev managed ClickHouse?
A dedicated ClickHouse instance (24.3, 24.8 or 25.1) on AWS that Selfhost.dev provisions and runs for you: a columnar OLAP database for real-time analytics, with optional ClickHouse Keeper high availability, HTTP and native protocols, SQL users, automated backups and snapshots, forking, live metrics with alerts, encryption at rest and BYOC. It is our first analytics engine, alongside managed PostgreSQL, MySQL and Redis, with the same pay-as-you-go pricing.
What is ClickHouse good for and what is it not?
ClickHouse is an OLAP engine built for analytics: fast aggregations over huge volumes of events, logs, metrics and time-series data, powering dashboards and reporting in real time. It is not a transactional (OLTP) database, so it does not replace PostgreSQL or MySQL for your app's reads and writes. A common pattern is to run both: a relational database for the app and ClickHouse for the analytics on top.
How does high availability work?
Multi-AZ ClickHouse runs a 3-node cluster with embedded ClickHouse Keeper (a Raft quorum) and promotes a replica automatically if the primary fails. Failover is best-effort, optimized for minimal data loss rather than a synchronous zero-data-loss guarantee. Distributed queries run across the selfhost_cluster and all nodes accept reads.
How do I connect to it?
Two protocols are exposed: the HTTP API on port 8123 (used by JDBC, most drivers and clickhouse-client by default) and the native TCP protocol on port 9000 for high-throughput clients. You authenticate as a SQL user and connect from any ClickHouse-compatible client, BI tool or your own code.
Is this a managed ClickHouse Cloud alternative?
Yes. You get a dedicated instance you size yourself, with live metrics and logs you can see into, pay-as-you-go by the hour with no tier minimums and the option to run it in your own AWS account with BYOC. You also manage it from your AI editor over MCP. It is a self-hosted-style ClickHouse you control, run for you.
Which ClickHouse versions are supported?
ClickHouse 24.3, 24.8 and 25.1. New instances default to 24.8, the LTS release and you can pick another version in the creation wizard if a workload needs it.
Can I run managed ClickHouse in my own AWS account?
Yes. BYOC runs the instance inside your own AWS account, connected via access keys or an IAM role, while Selfhost.dev handles provisioning, backups, failover and upgrades. You pay AWS directly for the compute and storage and our fee is 70% of the managed price. You keep ownership of the infrastructure and the data residency and if you ever leave, the database stays in your account.
How much does managed ClickHouse cost?
Pay-as-you-go, metered by the hour against prepaid credits. ClickHouse instances start at 4 GB RAM, so the smallest starts at about $27/mo in ap-south-1 (Mumbai), plus about $5/mo for the public IPv4 added at deploy and analytics workloads simply pick a larger instance as needed. There are no tiers or subscriptions and everything pauses at a zero balance. Build your exact bill.
How do I load or migrate data in?
It is standard ClickHouse, so standard paths work: INSERT over HTTP or native, clickhouse-client and the s3 and remote table functions to pull from existing sources. Stream events in from your app or pipeline, nothing proprietary to convert and nothing proprietary holding you here later. Moving from Postgres is the common case: a one-time bulk load, ClickHouse's native postgresql() table function or continuous CDC replication, depending on your data size and whether the sync has to be live. Our guide walks through all five ways to move data from Postgres into ClickHouse.

Real-time analytics, run for you.
Live in minutes.

Deploy ClickHouse

Your $5 welcome credit runs the smallest ClickHouse instance, no card needed.